Info

feed rate

Fig. 3.10. Comparison of experimental conversion values with the simulation using a reaction rate constant of 100% (solid line) and 12% (dotted line) ([18], reprinted from Chem. Eng. Sci., Vol 57, Beckmann et al., Pages 1525-1530, Copyright 2002, with permission from Elsevier Science)

feed rate

Fig. 3.10. Comparison of experimental conversion values with the simulation using a reaction rate constant of 100% (solid line) and 12% (dotted line) ([18], reprinted from Chem. Eng. Sci., Vol 57, Beckmann et al., Pages 1525-1530, Copyright 2002, with permission from Elsevier Science)

kinetic and laboratory scale experiments. This column was applied to perform a set of experiments by varying the feed rate, the feed position (below, within and above the reactive section), the reflux ratio (2-15) and the pressure (500-650 kPa). The temperature in the rectifying section was controlled by varying the reboiler load. As a result, the overhead product was always the isobutene-methanol azeotrope containing small amounts of dimethyl ether formed in a side reaction.

Simulating the pilot column with the same model that was successfully used for the 80 mm column, the computed conversion was systematically too high (Fig. 3.10). Reaction kinetics (reduced reaction rates due to incomplete catalyst wetting, mass-transfer limitations, or maldistribution) and separation efficiency of the reactive packing were considered as possible reasons for these deviations. It was found that the sensitivity of the computed conversion against in these factors was small compared with the experimentally observed trends. The best agreement was attained by significant reduction of the reaction rate constant. Assuming a reduced reaction rate of 12%, the performance of the pilot column could be described satisfactorily as depicted in Fig. 3.10. This implies that pilot-scale experiments provide the only basis for scale-up. This conclusion is supported by the data shown in Fig. 3.11. In this figure, experimental conversion data is plotted against the feed location and compared with the model prediction using the reduced reaction rate of 12 %. For the upper feed positions, the agreement is good, the trend in the conversion for the lower feed positions is qualitatively correct.

However, the observed reduction in reaction rate still remains unexplained. The pilot-scale RD column was operated at similar space velocity and hydrodynamic

60 | T. Frey, F. Nierlich, T. Popken, D. Reusch, J. Stichlmair, and A. Tuchlenski 1

0 0

Post a comment